Objective

Set up an end-to-end pipeline using n8n on Windows to:

  1. Ingest documents (e.g., from Google Drive),
  2. Generate embeddings using Ollama,
  3. Store embeddings in Supabase,
  4. Retrieve relevant information from these embeddings,
  5. Answer user queries using DeepSeek-R1 (OpenRouter).

Prerequisites

  • Windows 10/11
  • WSL (Windows Subsystem for Linux) 
  • Git
  • Node.js (LTS)
  • Docker Desktop (for Ollama, if using Docker mode)
  • Supabase account
  • OpenRouter account with access to DeepSeek-R1

Architecture 

The architecture illustrates an automated document processing QnA chatbot built using N8N. The workflow starts with a trigger that initiates file download from Google Drive. The downloaded document is loaded using a default data loader and split into chunks with a Recursive Character Text Splitter. These chunks are then converted into vector embeddings using the Ollama Embeddings model and stored in a Supabase Vector Store. When a user sends a message via the chat interface, it triggers the Question and Answer Chain. This chain uses a Vector Store Retriever to fetch relevant document embeddings from Supabase and processes the query with an OpenRouter chat model to generate a response. The system is designed to fully automate QnA by integrating document ingestion, embedding, storage, retrieval, and response generation.

Architecture 

Step 1: Set Up n8n on Windows

Option 1: Native (using Node.js)

npm install n8n -g

n8n

Option 2: Using Docker

docker run -it –rm \

  -p 5678:5678 \

  -v ~/.n8n:/home/node/.n8n \

  n8nio/n8n

Access http://localhost:5678 in your browser.

Step 2: Set Up Ollama on Windows

Download Ollama

From the official site: https://ollama.com

Pull Model

ollama run all-minilm

This lightweight model is suitable for generating embeddings in your n8n pipeline.

Step 3: Create Supabase Vector Store

  1. Create a Supabase Project
  2. Create a table document_embeddings with columns:
  • id (uuid)
  • embedding (vector)
  • content (text)
  • metadata (jsonb)
  • created_at (timestamp)
  1. Enable pgvector extension:

create extension if not exists vector;

Step 4: Workflow 1 – Document Embedding

Nodes Flow:

  1. Trigger – Manual (when clicking “Test Workflow”)
  2. Google Drive – Downloads selected .docx or .pdf
  3. Default Data Loader – Extracts raw text
  4. Recursive Character Text Splitter – Breaks into chunks
  5. Embeddings (Ollama) – Generates embeddings
  6. Supabase Vector Store – Stores them in document_embeddings

Step 5: Workflow 2 – Q&A Chatbot

Nodes Flow:

  1. Trigger – On chat message received
  2. Supabase Vector Store Retriever – Retrieves matching documents based on embedding similarity
  3. OpenRouter Chat Model (DeepSeek-R1) – Generates human-like answers
  4. Question and Answer Chain – Combines documents + user query for model
  5. Response – Returns JSON like:

{ “response”: “Your summarized answer…” }

Example in Action

User Prompt:

tell me project overview in short

Model Output:

The project involves integrating ChirpStack, InfluxDB, and Grafana… real-time monitoring dashboards… key sensors monitored include AM319, VS350…

Outcome

We have now a functioning document-based Q&A system using:

  •  Documents from Google Drive
  •  Embeddings via Ollama
  •  Storage via Supabase
  •  Responses from DeepSeek-R1 (OpenRouter)

This enables users to ask questions and get intelligent answers grounded in uploaded documents.

RELATED ARTICLESMORE FROM AUTHOR